Databricks Machine Learning Professional Question 174
Single answerA retail company wants to implement a machine learning model in production that predicts whether a customer will make a purchase based on their real-time browsing behavior. The company receives a continuous stream of browsing data from its website, and predictions must be made in near real-time. Which Databricks feature should the company use to implement this solution?
- A
Databricks Structured Streaming to apply the model for continuous inference
- B
Databricks Jobs to schedule periodic batch inference
- C
Databricks AutoML to automatically train and deploy the model
- D
Databricks Delta Live Tables to store the predictions for later analysis
Show answer and explanation
Correct answer: A
Explanation
For real-time predictions on continuously incoming data, Databricks Structured Streaming is the most appropriate feature. It allows for streaming data to be processed in real-time and enables applying machine learning models for continuous inference. Other options, like Jobs or Delta Live Tables, are not designed for real-time inference scenarios.
- A. Correct.
Correct: Databricks Structured Streaming is designed for processing continuous streams of data and can be used for real-time inference by applying a trained machine learning model to incoming data.
- B. Incorrect.
Incorrect: Databricks Jobs are better suited for batch processing and scheduled inference tasks, not for real-time, continuous inference on streaming data.
- C. Incorrect.
Incorrect: Databricks AutoML is helpful for training and automating model development, but it does not directly enable real-time inference on streaming data.
- D. Incorrect.
Incorrect: Databricks Delta Live Tables are used for data engineering tasks such as building ETL pipelines and maintaining data quality, not for applying machine learning models in real-time.